# Efficient Multimodal Streaming Recommendation via Expandable Side Mixture-of-Experts

arXiv: https://arxiv.org/abs/2508.05993 · HTML: https://mlsanity.com/paper/2508.05993/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Placed in MoE for recommender systems, one of 12 directions in Mixture-of-experts.
- Builds on 42 of its 74 references that are in the map.
- First posted on arXiv 8 Aug 2025.

## Placed in

- Mixture-of-experts / [MoE for recommender systems](https://mlsanity.com/moe/moe-for-recommender-systems/)

## Builds on (most cited first, 30 of 42)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Adam: A Method for Stochastic Optimization](https://mlsanity.com/paper/1412.6980/) (2014)
- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [Overcoming catastrophic forgetting in neural networks](https://mlsanity.com/paper/1612.00796/) (2016)
- [Parameter-Efficient Transfer Learning for NLP](https://mlsanity.com/paper/1902.00751/) (2019)
- [Overcoming catastrophic forgetting with hard attention to the task](https://mlsanity.com/paper/1801.01423/) (2018)
- [Learning to Learn without Forgetting by Maximizing Transfer and Minimizing Interference](https://mlsanity.com/paper/1810.11910/) (2018)
- [Variational Continual Learning](https://mlsanity.com/paper/1710.10628/) (2017)
- [AlphaEdit: Null-Space Constrained Knowledge Editing for Language Models](https://mlsanity.com/paper/2410.02355/) (2024)
- [LST: Ladder Side-Tuning for Parameter and Memory Efficient Transfer Learning](https://mlsanity.com/paper/2206.06522/) (2022)
- [Boosting Continual Learning of Vision-Language Models via Mixture-of-Experts Adapters](https://mlsanity.com/paper/2403.11549/) (2024)
- [Text Is All You Need: Learning Language Representations for Sequential Recommendation](https://mlsanity.com/paper/2305.13731/) (2023)
- [Uncertainty-based Continual Learning with Adaptive Regularization](https://mlsanity.com/paper/1905.11614/) (2019)
- [A Neural Dirichlet Process Mixture Model for Task-Free Continual Learning](https://mlsanity.com/paper/2001.00689/) (2020)
- [Where to Go Next for Recommender Systems? ID- vs. Modality-based Recommender Models Revisited](https://mlsanity.com/paper/2303.13835/) (2023)
- [A Content-Driven Micro-Video Recommendation Dataset at Scale](https://mlsanity.com/paper/2309.15379/) (2023)
- [How to Retrain Recommender System? A Sequential Meta-Learning Method](https://mlsanity.com/paper/2005.13258/) (2020)
- [GraphSAIL: Graph Structure Aware Incremental Learning for Recommender Systems](https://mlsanity.com/paper/2008.13517/) (2020)
- [LGMRec: Local and Global Graph Learning for Multimodal Recommendation](https://mlsanity.com/paper/2312.16400/) (2023)
- [Multimodal Recommender Systems: A Survey](https://mlsanity.com/paper/2302.03883/) (2023)
- [IISAN: Efficiently Adapting Multimodal Representation for Sequential Recommendation with Decoupled PEFT](https://mlsanity.com/paper/2404.02059/) (2024)
- [Interaction-level Membership Inference Attack Against Federated Recommender Systems](https://mlsanity.com/paper/2301.10964/) (2023)
- [ADER: Adaptively Distilled Exemplar Replay Towards Continual Learning for Session-based Recommendation](https://mlsanity.com/paper/2007.12000/) (2020)
- [DTL: Disentangled Transfer Learning for Visual Recognition](https://mlsanity.com/paper/2312.07856/) (2023)
- [Comprehensive Privacy Analysis on Federated Recommender System against Attribute Inference Attacks](https://mlsanity.com/paper/2205.11857/) (2022)
- [A Practical Incremental Method to Train Deep CTR Models](https://mlsanity.com/paper/2009.02147/) (2020)
- [Memory Augmented Neural Model for Incremental Session-based Recommendation](https://mlsanity.com/paper/2005.01573/) (2020)
- [A Survey on Incremental Update for Neural Recommender Systems](https://mlsanity.com/paper/2303.02851/) (2023)
- [Continuous Input Embedding Size Search For Recommender Systems](https://mlsanity.com/paper/2304.03501/) (2023)
- [ATLAS: Adapter-Based Multi-Modal Continual Learning with a Two-Stage Learning Strategy](https://mlsanity.com/paper/2410.10923/) (2024)

## Built on by (most cited first, 2 of 2)

- [Efficient Content-based Recommendation Model Training via Noise-aware Coreset Selection](https://mlsanity.com/paper/2601.10067/) (2026)
- [TimeRoute: Time-Aware Modality Routing and Diffusion for Multi-Modal Recommendation](https://mlsanity.com/paper/2608.10983/) (2026)

## Abstract (arXiv)

Streaming recommender systems (SRSs) are widely deployed in real-world applications, where user interests shift and new items arrive over time. As a result, effectively capturing users' latest preferences is challenging, as interactions reflecting recent interests are limited and new items often lack sufficient feedback. A common solution is to enrich item representations using multimodal encoders (e.g., BERT or ViT) to extract visual and textual features. However, these encoders are pretrained on general-purpose tasks: they are not tailored to user preference modeling, and they overlook the fact that user tastes toward modality-specific features such as visual styles and textual tones can also drift over time. This presents two key challenges in streaming scenarios: the high cost of fine-tuning large multimodal encoders, and the risk of forgetting long-term user preferences due to continuous model updates. To tackle these challenges, we propose Expandable Side Mixture-of-Experts (XSMoE), a memory-efficient framework for multimodal streaming recommendation. XSMoE attaches lightweight side-tuning modules consisting of expandable expert networks to frozen pretrained encoders and incrementally expands them in response to evolving user feedback. A gating router dynamically combines expert and backbone outputs, while a utilization-based pruning strategy maintains model compactness. By learning new patterns through expandable experts without overwriting previously acquired knowledge, XSMoE effectively captures both cold start and shifting preferences in multimodal features. Experiments on three real-world datasets demonstrate that XSMoE outperforms state-of-the-art baselines in both recommendation quality and computational efficiency.